search

dddraxxx / Weakly-Supervised-Camouflaged-Object-Detection-with-Scribble-Annotations

Code for the AAAI 2023 paper "Weakly-Supervised Camouflaged Object Detection with Scribble Annotations"
https://arxiv.org/abs/2207.14083
Other
51 stars 8 forks source link

Results on NC4K #3

Closed GewelsJI closed 1 year ago

GewelsJI commented 1 year ago

Hi, @dddraxxx

Could you please provide the prediction results on NC4K dataset?

Thanks.

dddraxxx commented 1 year ago

We provide the the code and model weight so you may test it in any dataset you want.

GewelsJI commented 1 year ago

Hi, @dddraxxx

Sorry to bother you again. I have a question about how to inference with your model.

When I run test.py, I got a bug here:

RuntimeError: Error(s) in loading state_dict for Net:
        Unexpected key(s) in state_dict: "rw.gamma", "rw.query_conv.weight", "rw.query_conv.bias", "rw.key_conv.weight", "rw.key_conv.bias", "rw.value_conv.weight", "rw.value_conv.bias", "ca.0.conv1.weight", "ca.0.conv1.bias", "ca.0.bn1.weight", "ca.0.bn1.bias", "ca.0.bn1.running_mean", "ca.0.bn1.running_var", "ca.0.bn1.num_batches_tracked", "ca.0.conv_h.weight", "ca.0.conv_h.bias", "ca.0.conv_w.weight", "ca.0.conv_w.bias", "ca.1.conv1.weight", "ca.1.conv1.bias", "ca.1.bn1.weight", "ca.1.bn1.bias", "ca.1.bn1.running_mean", "ca.1.bn1.running_var", "ca.1.bn1.num_batches_tracked", "ca.1.conv_h.weight", "ca.1.conv_h.bias", "ca.1.conv_w.weight", "ca.1.conv_w.bias", "ca.2.conv1.weight", "ca.2.conv1.bias", "ca.2.bn1.weight", "ca.2.bn1.bias", "ca.2.bn1.running_mean", "ca.2.bn1.running_var", "ca.2.bn1.num_batches_tracked", "ca.2.conv_h.weight", "ca.2.conv_h.bias", "ca.2.conv_w.weight", "ca.2.conv_w.bias", "ca.3.conv1.weight", "ca.3.conv1.bias", "ca.3.bn1.weight", "ca.3.bn1.bias", "ca.3.bn1.running_mean", "ca.3.bn1.running_var", "ca.3.bn1.num_batches_tracked", "ca.3.conv_h.weight", "ca.3.conv_h.bias", "ca.3.conv_w.weight", "ca.3.conv_w.bias", "ca.4.conv1.weight", "ca.4.conv1.bias", "ca.4.bn1.weight", "ca.4.bn1.bias", "ca.4.bn1.running_mean", "ca.4.bn1.running_var", "ca.4.bn1.num_batches_tracked", "ca.4.conv_h.weight", "ca.4.conv_h.bias", "ca.4.conv_w.weight", "ca.4.conv_w.bias", "edge_extract.0.weight", "edge_extract.0.bias", "edge_extract.1.weight", "edge_extract.1.bias", "edge_extract.1.running_mean", "edge_extract.1.running_var", "edge_extract.1.num_batches_tracked", "edge_extract.3.weight", "edge_extract.3.bias", "edge_extract.4.weight", "edge_extract.4.bias", "edge_extract.4.running_mean", "edge_extract.4.running_var", "edge_extract.4.num_batches_tracked".

I further check the net.py, and I also find some extra unused modules such as edge_head. Could you help me on this problem?

Thanks Ge-Peng.

dddraxxx commented 1 year ago

Hi, @dddraxxx

Sorry to bother you again. I have a question about how to inference with your model.

When I run test.py, I got a bug here:

RuntimeError: Error(s) in loading state_dict for Net:
        Unexpected key(s) in state_dict: "rw.gamma", "rw.query_conv.weight", "rw.query_conv.bias", "rw.key_conv.weight", "rw.key_conv.bias", "rw.value_conv.weight", "rw.value_conv.bias", "ca.0.conv1.weight", "ca.0.conv1.bias", "ca.0.bn1.weight", "ca.0.bn1.bias", "ca.0.bn1.running_mean", "ca.0.bn1.running_var", "ca.0.bn1.num_batches_tracked", "ca.0.conv_h.weight", "ca.0.conv_h.bias", "ca.0.conv_w.weight", "ca.0.conv_w.bias", "ca.1.conv1.weight", "ca.1.conv1.bias", "ca.1.bn1.weight", "ca.1.bn1.bias", "ca.1.bn1.running_mean", "ca.1.bn1.running_var", "ca.1.bn1.num_batches_tracked", "ca.1.conv_h.weight", "ca.1.conv_h.bias", "ca.1.conv_w.weight", "ca.1.conv_w.bias", "ca.2.conv1.weight", "ca.2.conv1.bias", "ca.2.bn1.weight", "ca.2.bn1.bias", "ca.2.bn1.running_mean", "ca.2.bn1.running_var", "ca.2.bn1.num_batches_tracked", "ca.2.conv_h.weight", "ca.2.conv_h.bias", "ca.2.conv_w.weight", "ca.2.conv_w.bias", "ca.3.conv1.weight", "ca.3.conv1.bias", "ca.3.bn1.weight", "ca.3.bn1.bias", "ca.3.bn1.running_mean", "ca.3.bn1.running_var", "ca.3.bn1.num_batches_tracked", "ca.3.conv_h.weight", "ca.3.conv_h.bias", "ca.3.conv_w.weight", "ca.3.conv_w.bias", "ca.4.conv1.weight", "ca.4.conv1.bias", "ca.4.bn1.weight", "ca.4.bn1.bias", "ca.4.bn1.running_mean", "ca.4.bn1.running_var", "ca.4.bn1.num_batches_tracked", "ca.4.conv_h.weight", "ca.4.conv_h.bias", "ca.4.conv_w.weight", "ca.4.conv_w.bias", "edge_extract.0.weight", "edge_extract.0.bias", "edge_extract.1.weight", "edge_extract.1.bias", "edge_extract.1.running_mean", "edge_extract.1.running_var", "edge_extract.1.num_batches_tracked", "edge_extract.3.weight", "edge_extract.3.bias", "edge_extract.4.weight", "edge_extract.4.bias", "edge_extract.4.running_mean", "edge_extract.4.running_var", "edge_extract.4.num_batches_tracked".

I further check the net.py, and I also find some extra unused modules such as edge_head. Could you help me on this problem?

Thanks Ge-Peng.

Sorry for the mistake. I have uploaded the code and a correct version of model_weight. Please use it instead.

GewelsJI commented 1 year ago

Thanks. But how to make a txt list for testing?

https://github.com/dddraxxx/Weakly-Supervised-Camouflaged-Object-Detection-with-Scribble-Annotations/blob/main/data/dataset.py#L76

dddraxxx commented 1 year ago

Thanks. But how to make a txt list for testing?

https://github.com/dddraxxx/Weakly-Supervised-Camouflaged-Object-Detection-with-Scribble-Annotations/blob/main/data/dataset.py#L76

You can refer to this in #2